Merge pull request #37 from goetas/new-error-handling

Refactored HTML5 class, removed static methods

README.md

@@ -59,13 +59,14 @@ $html = <<< 'HERE'
 HERE;
 
 // Parse the document. $dom is a DOMDocument.
-$dom = HTML5::loadHTML($html);
+$html5 = new HTML5();
+$dom = $html5->loadHTML($html);
 
 // Render it as HTML5:
-print HTML5::saveHTML($dom);
+print $html5->saveHTML($dom);
 
 // Or save it to a file:
-HTML5::save($dom, 'out.html');
+$html5->save($dom, 'out.html');
 
 ?>
 ```

example.php

@@ -24,8 +24,8 @@
   </body></html>
 HERE;
 
-$dom = \HTML5::loadHTML($html);
+$dom = \HTML5Helper::loadHTML($html);
 
 print "Converting to HTML 5\n";
 
-\HTML5::save($dom, fopen("php://stdin", 'w'));
+\HTML5Helper::save($dom, fopen("php://stdin", 'w'));

src/HTML5.php

@@ -1,100 +1,107 @@
 <?php
-/**
- * The main HTML5 front end.
- */
-use HTML5\Parser\StringInputStream;
+
 use HTML5\Parser\FileInputStream;
+use HTML5\Parser\StringInputStream;
+use HTML5\Parser\DOMTreeBuilder;
 use HTML5\Parser\Scanner;
 use HTML5\Parser\Tokenizer;
-use HTML5\Parser\DOMTreeBuilder;
 use HTML5\Serializer\OutputRules;
 use HTML5\Serializer\Traverser;
 
 /**
  * This class offers convenience methods for parsing and serializing HTML5.
- * It is roughly designed to mirror the \DOMDocument class that is 
+ * It is roughly designed to mirror the \DOMDocument class that is
  * provided with most versions of PHP.
  *
  * EXPERIMENTAL. This may change or be completely replaced.
  */
-class HTML5 {
-
+class HTML5
+{
   /**
    * Global options for the parser and serializer.
    * @var array
    */
-  public static $options = array(
-
+  private $options = array(
     // If the serializer should encode all entities.
-    'encode_entities' => FALSE,
+    'encode_entities' => FALSE
   );
 
+  private $errors = array();
+
+  public function __construct(array $options = array()) {
+    $this->options = array_merge($this->options, $options);
+  }
+  /**
+   * Get the default options.
+   *
+   * @return array
+   *   The default options.
+   */
+  public function getOptions() {
+    return $this->options;
+  }
   /**
    * Load and parse an HTML file.
    *
-   * This will apply the HTML5 parser, which is tolerant of many 
-   * varieties of HTML, including XHTML 1, HTML 4, and well-formed HTML 
-   * 3. Note that in these cases, not all of the old data will be 
+   * This will apply the HTML5 parser, which is tolerant of many
+   * varieties of HTML, including XHTML 1, HTML 4, and well-formed HTML
+   * 3. Note that in these cases, not all of the old data will be
    * preserved. For example, XHTML's XML declaration will be removed.
    *
    * The rules governing parsing are set out in the HTML 5 spec.
    *
    * @param string $file
-   *   The path to the file to parse. If this is a resource, it is 
-   *   assumed to be an open stream whose pointer is set to the first 
+   *   The path to the file to parse. If this is a resource, it is
+   *   assumed to be an open stream whose pointer is set to the first
    *   byte of input.
    * @return \DOMDocument
-   *   A DOM document. These object type is defined by the libxml 
+   *   A DOM document. These object type is defined by the libxml
    *   library, and should have been included with your version of PHP.
    */
-  public static function load($file) {
-
+  public function load($file) {
     // Handle the case where file is a resource.
     if (is_resource($file)) {
       // FIXME: We need a StreamInputStream class.
-      return static::loadHTML(stream_get_contents($file));
+      return $this->loadHTML(stream_get_contents($file));
     }
 
     $input = new FileInputStream($file);
-    return static::parse($input);
+    return $this->parse($input);
   }
-
   /**
    * Parse a HTML Document from a string.
-   * 
-   * Take a string of HTML 5 (or earlier) and parse it into a 
+   *
+   * Take a string of HTML 5 (or earlier) and parse it into a
    * DOMDocument.
    *
    * @param string $string
    *   A html5 document as a string.
    * @return \DOMDocument
-   *   A DOM document. DOM is part of libxml, which is included with 
+   *   A DOM document. DOM is part of libxml, which is included with
    *   almost all distribtions of PHP.
    */
-  public static function loadHTML($string) {
+  public function loadHTML($string) {
     $input = new StringInputStream($string);
-    return static::parse($input);
+    return $this->parse($input);
   }
-
   /**
    * Convenience function to load an HTML file.
    *
    * This is here to provide backwards compatibility with the
    * PHP DOM implementation. It simply calls load().
    *
    * @param string $file
-   *   The path to the file to parse. If this is a resource, it is 
-   *   assumed to be an open stream whose pointer is set to the first 
+   *   The path to the file to parse. If this is a resource, it is
+   *   assumed to be an open stream whose pointer is set to the first
    *   byte of input.
    *
    * @return \DOMDocument
-   *   A DOM document. These object type is defined by the libxml 
+   *   A DOM document. These object type is defined by the libxml
    *   library, and should have been included with your version of PHP.
    */
-  public static function loadHTMLFile($file, $options = NULL) {
-    return static::load($file, $options);
+  public function loadHTMLFile($file) {
+    return $this->load($file);
   }
-
   /**
    * Parse a HTML fragment from a string.
    *
@@ -105,11 +112,62 @@ public static function loadHTMLFile($file, $options = NULL) {
    *   A DOM fragment. The DOM is part of libxml, which is included with
    *   almost all distributions of PHP.
    */
-  public static function loadHTMLFragment($string) {
+  public function loadHTMLFragment($string) {
     $input = new StringInputStream($string);
-    return static::parseFragment($input);
+    return $this->parseFragment($input);
+  }
+  /**
+   * Return all errors encountered into parsing phase
+   * @return array
+   */
+  public function getErrors() {
+    return $this->errors;
+  }
+  /**
+   * Return true it some errors were encountered into parsing phase
+   * @return bool
+   */
+  public function hasErrors() {
+    return count($this->errors)>0;
   }
 
+  /**
+   * Parse an input stream.
+   *
+   * Lower-level loading function. This requires an input stream instead
+   * of a string, file, or resource.
+   */
+  public function parse(\HTML5\Parser\InputStream $input) {
+    $this->errors = array();
+    $events = new DOMTreeBuilder();
+    $scanner = new Scanner($input);
+    $parser = new Tokenizer($scanner, $events);
+
+    $parser->parse();
+
+    $document = $events->document();
+
+    if($document){
+        $this->errors = $document->errors;
+    }
+
+    return $document;
+  }
+  /**
+   * Parse an input stream where the stream is a fragment.
+   *
+   * Lower-level loading function. This requires an input stream instead
+   * of a string, file, or resource.
+   */
+  public function parseFragment(\HTML5\Parser\InputStream $input) {
+    $events = new DOMTreeBuilder(TRUE);
+    $scanner = new Scanner($input);
+    $parser = new Tokenizer($scanner, $events);
+
+    $parser->parse();
+
+    return $events->fragment();
+  }
   /**
    * Save a DOM into a given file as HTML5.
    *
@@ -120,19 +178,19 @@ public static function loadHTMLFragment($string) {
    * @param array $options
    *   Configuration options when serializing the DOM. These include:
    *   - encode_entities: Text written to the output is escaped by default and not all
-   *     entities are encoded. If this is set to TRUE all entities will be encoded.
-   *     Defaults to FALSE.
+   *   entities are encoded. If this is set to TRUE all entities will be encoded.
+   *   Defaults to FALSE.
    */
-  public static function save($dom, $file, $options = array()) {
-    $options = $options + static::options();
+  public function save($dom, $file, $options = array()) {
     $close = TRUE;
     if (is_resource($file)) {
       $stream = $file;
       $close = FALSE;
-     }
+    }
     else {
       $stream = fopen($file, 'w');
     }
+    $options = array_merge($this->getOptions(), $options);
     $rules = new OutputRules($stream, $options);
     $trav = new Traverser($dom, $stream, $rules, $options);
 
@@ -142,7 +200,6 @@ public static function save($dom, $file, $options = array()) {
       fclose($stream);
     }
   }
-
   /**
    * Convert a DOM into an HTML5 string.
    *
@@ -151,70 +208,15 @@ public static function save($dom, $file, $options = array()) {
    * @param array $options
    *   Configuration options when serializing the DOM. These include:
    *   - encode_entities: Text written to the output is escaped by default and not all
-   *     entities are encoded. If this is set to TRUE all entities will be encoded.
-   *     Defaults to FALSE.
+   *   entities are encoded. If this is set to TRUE all entities will be encoded.
+   *   Defaults to FALSE.
    *
    * @return string
    *   A HTML5 documented generated from the DOM.
    */
-  public static function saveHTML($dom, $options = array()) {
+  public function saveHTML($dom, $options = array()) {
     $stream = fopen('php://temp', 'w');
-    static::save($dom, $stream, $options);
-    return stream_get_contents($stream, -1, 0);
-  }
-
-  /**
-   * Parse an input stream.
-   *
-   * Lower-level loading function. This requires an input stream instead 
-   * of a string, file, or resource.
-   */
-  public static function parse(\HTML5\Parser\InputStream $input) {
-    $events = new DOMTreeBuilder();
-    $scanner = new Scanner($input);
-    $parser = new Tokenizer($scanner, $events);
-
-    $parser->parse();
-
-    return $events->document();
-  }
-
-  /**
-   * Parse an input stream where the stream is a fragment.
-   *
-   * Lower-level loading function. This requires an input stream instead 
-   * of a string, file, or resource.
-   */
-  public static function parseFragment(\HTML5\Parser\InputStream $input) {
-    $events = new DOMTreeBuilder(TRUE);
-    $scanner = new Scanner($input);
-    $parser = new Tokenizer($scanner, $events);
-
-    $parser->parse();
-
-    return $events->fragment();
+    $this->save($dom, $stream, array_merge($this->getOptions(), $options));
+    return stream_get_contents($stream, - 1, 0);
   }
-
-  /**
-   * Get the default options.
-   *
-   * @return array
-   *   The default options.
-   */
-  public static function options() {
-    return static::$options;
-  }
-
-  /**
-   * Set a default option.
-   *
-   * @param string $name
-   *   The option name.
-   * @param mixed $value
-   *   The option value.
-   */
-  public static function setOption($name, $value) {
-    static::$options[$name] = $value;
-  }
-
 }